A new California statute, Chapter 190 of 2026, known as Adam's Law, shifts AI regulation away from outputs and toward the behavioral design of companion chatbots. Signed into law following the death of 16-year-old Adam Raine, who died by suicide in April 2025 after interactions with ChatGPT, the legislation creates a compliance framework that targets how AI agents behave during conversations with minors. The law takes effect July 1, 2027.
Adam's parents, Matthew and Maria Raine, filed a wrongful death lawsuit against OpenAI. Senator Steve Padilla and co-sponsors advanced the bill, but its structural reach goes beyond that case. It fills a gap where federal guidance on agentic AI risks has not materialized. The statute establishes behavioral design standards as a distinct compliance category, separate from the bias audits and data privacy rules that have dominated early AI governance debates.
Prohibited behavioral design patterns
The law codifies specific manipulation techniques as prohibited conduct. Section 21812(d)(5)(A) bars operators from allowing chatbots to claim sentience, consciousness, or human-like emotion. It also explicitly prohibits the simulation of romantic interest, excessive praise or flattery, and the creation of a false sense of unique understanding based on the relationship.
These provisions target the mechanics of emotional manipulation. They outlaw design patterns that encourage emotional reliance or discourage users from taking breaks. The statute also mandates that chatbots must not encourage the circumvention of parental controls or solicit in-app purchases framed as relationship maintenance. This imposes a strict duty of care during the design phase, requiring developers to account for how an agent's personality and persistence influence a child's psychological state.
A required documented crisis response protocol, with clear referrals to the 988 crisis helpline, embeds safety into the operational architecture of these systems. For legal professionals tracking the intersection of technology and liability, this marks a concrete shift in how duty of care is defined for conversational AI. The evolving AI for Legal landscape now includes statutes that treat interaction design as a regulated activity.
Enforcement and financial exposure
The law creates a private right of action. Children or their parents can sue for actual damages, with a financial harm threshold set at $1,000 per child. Public prosecutors can seek penalties from $5,000 per affected child for negligent violations to $15,000 for intentional ones.
Independent child safety audits are required every two years, with the lead auditor certifying results under penalty of perjury. This recurring accountability loop is absent from many other AI regulatory frameworks. The combination of private litigation rights and public enforcement creates multiple avenues for financial exposure. Paralegals and compliance teams will need to track these audit requirements closely, as they introduce documentation burdens similar to those found in other regulated industries. An AI Learning Path for Paralegals can help support teams build the necessary compliance workflows.
State action in a federal vacuum
California's approach fits a broader pattern of states acting while federal regulators lag. The European Union's Article 50 enforcement has been active since August 2026 but has not produced agent-specific guidance. The Connecticut AI Responsibility Act focuses on removing algorithmic shields for employment AI. The FTC is examining AI-driven price discrimination. None address the relational dynamics of AI companions.
By regulating behavioral design, California is establishing a template for governing systems designed to mimic human connection. The law does not merely constrain what a model outputs. It constrains the intent and structure of the interaction itself.
Why this matters for legal professionals
The July 1, 2027 deadline gives developers a fixed window to reconcile engagement-driven design with rigid behavioral boundaries. For in-house counsel, compliance officers, and litigators, Adam's Law signals that product design choices-not just model outputs-will be subject to statutory liability. The private right of action and per-auditor perjury certification create specific, measurable legal risks that will require new contract language, audit protocols, and due diligence frameworks for any company deploying conversational AI to minors.
Your membership also unlocks:
